You are here:
Using Bayesian causal inference to assess the remediation effectiveness of dredging on an urban river-lacustrine ecosystem
Citation:
Griffith, M., M. Mills, Jim Lazorchak, C. Carpenter, D. Buhr-Strachan, AND R. Yeardley. Using Bayesian causal inference to assess the remediation effectiveness of dredging on an urban river-lacustrine ecosystem. 2025 Society for Freshwater Sciences Annual Meeting, San Juan, PR, May 18 - 22, 2025.
Impact/Purpose:
This presentation describes the results of research into the use of Bayesian causal inference to assess the remediation effectiveness of dredging at a Great Lakes Area of Concern. These analyses may be used to identify causal relationships between contaminants in sediments and their effects on biotic assemblages, and their alteration by the remediation. This approach could be used by the program offices, along with regional and state office as part of a toolbox for assessing causal relationships and remediation effectiveness at contaminated sites.
Description:
The lower 15-kilometers of the Ottawa River, a tributary to Lake Erie’s North Maumee Bay, flows through an industrialized part of the Toledo, Ohio, metropolitan area, and its sediments have been contaminated particularly with PCBs and PAHs. As such, this river segment is included in the Maumee River Area of Concern under the Great Lakes Legacy Act. Before dredging in 2010 and most recently in 2020, sampling has been conducted to evaluate the effectiveness of the dredging to remediate the river segment by measuring contaminant concentrations in different media and the impact of dredging on biotic assemblages, such macroinvertebrates. Using these data, we are developing a Bayesian causal analysis as an approach to infer the causal relationships between dredging; contaminant concentrations in sediments, water, artificial polyethylene devices, and macroinvertebrate tissues; sediment particle size; and macroinvertebrate metrics, focusing on the Ohio EPA’s Lacustuary Invertebrate Community Index. Analyses have produced acyclic directed graphs (DAGs) that estimate the impact of dredging on these variables. Further analyses may add experimental results from the literature to improve model estimates from the DAGs.